|Publication number||US5561477 A|
|Application number||US 08/329,553|
|Publication date||Oct 1, 1996|
|Filing date||Oct 26, 1994|
|Priority date||Oct 26, 1994|
|Also published as||CN1098599C, CN1137724A, DE69521966D1, DE69521966T2, EP0710031A2, EP0710031A3, EP0710031B1|
|Publication number||08329553, 329553, US 5561477 A, US 5561477A, US-A-5561477, US5561477 A, US5561477A|
|Inventors||Peter P. Polit|
|Original Assignee||Thomson Consumer Electronics, Inc.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (6), Non-Patent Citations (1), Referenced by (77), Classifications (29), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention is related to the field of digital image signal processing, and more particularly to a system for coding a television signal in the presence of an image intensity gradient such as a luminance gradient associated with image fading, e.g., from a normal image to black level or vice-versa.
A video sequence such as represented by a television signal, for example, is a series of motionless images shown in rapid succession to give a viewer an impression of continuous motion. Each frame conveys distinctive image information, and the high frame rate necessary to achieve an appearance of motion often results in a significant amount of redundant temporal information among adjacent frames. Motion compensation coding, a form of data compression, is a form of predictive coding in a temporal dimension that is often used to attempt to remove such temporal redundancy.
In the absence of a scene change from one frame to the next, image motion from one frame to the next accounts for much of the variation in intensity from one frame to the next. In motion compensated predictive image coding, the current frame is predicted from a previously coded frame by estimating the motion between the two frames and compensating for the motion. The difference between the current frame and the prediction of the current frame is commonly called the (motion compensated) residual image, which is coded. The energy in the residual image is typically much less than the energy in the original image due to the removal of the redundant temporal information. Encoding the residual information rather than the original image information results in a more efficient use of the data bits available for coding.
Motion compensated predictive coding may assume many forms. One popular approach is based on block matching. In this approach the current image frame is partitioned into a prescribed number of rectangular regions or blocks, and a search is performed for the displacement in an adjacent frame which produces the best match among possible blocks in the adjacent frame. A motion vector with associated x,y coordinates establishes the relationship between the block in the current frame and its best match in the adjacent frame. Motion compensated (residual image) coding is an example of inter-frame (frame-to-frame) coding. In cases where motion compensated residual coding does not produce acceptable results (e.g., where the prediction is not good as when a scene changes from one frame to the next), better results may be obtained by intra-frame coding, where the video information of the frame is itself coded without motion compensation.
Various types of image coding/compression, including inter-frame predictive motion compensated coding as discussed above, are discussed by Ang et al. in Video Compression Makes Big Gains, IEEE Spectrum, October 1991, for example. In particular, this article describes a CCITT H.261 video motion coder including provision for intra-frame coding, and inter-frame predictive motion compensated residual coding compatible with the proposed MPEG (Moving Pictures Expert Group) image coding standard (ISO/IEC 13818-2, November 1993). The proposed MPEG standard also employs a motion compression algorithm with both inter- and intra-frame modes.
The present inventor has recognized that under certain conditions some motion estimators are inefficient enough to produce an image "blockiness" artifact (a noticeable difference in detail among pixel blocks that constitute an image). Such an artifact may be produced, for example, by motion estimators employing Mean Square Error or Mean Absolute Error processing, due to an erroneous match. More specifically, it has been recognized that, during an image fade from a normal image to black or vice-versa, frame-to-frame changes in image texture (detail) cannot be accurately tracked by other than highly sophisticated motion estimators. This inability to track frame-to-frame image detail leads to luminance tracking instead. Luminance tracking produces wrong information from the motion estimator, since if the image is motionless a luminance change with fading will falsely suggest motion. In other words, a frame-to-frame luminance gradient associated with image fading may fool the motion estimator into thinking there is motion when the motion estimator cannot discern that the image detail has not changed. Such tracking of a luminance gradient produces random, unpredictable results (e.g., "motion vectors") which are essentially useless and detrimental to coding efficiency.
In accordance with the principles of the present invention, it is herein recognized as beneficial in a video signal processing system including motion compensation to assign a value of zero to motion vectors under certain conditions. One of these conditions has been described above in the context of an image intensity gradient such as is commonly associated with image fading, where it has been found that zeroing motion vectors significantly reduces or eliminates the described "blockiness" artifacts. Zeroing motion vectors is also beneficial in situations where a motion estimator improperly detects motion. This may occur when a motion estimator does not perform a specific check for a (0,0) vector for a still image, or for still regions of an image frame.
A disclosed system zeroes motion vectors when a scene intensity change is detected, e.g., image fading from normal to black or vice-versa, and fading from one normal scene to another. For this purpose a sequence of three adjacent image frames is evaluated. Each frame is partitioned into a similar arrangement of segments. The sum of the absolute values of differences between similarly located pixels in segments from adjacent frames is obtained for each of the two pairs of adjacent frame segments. The ratio of the sums of the absolute pixel differences is obtained. Fading is indicated if the ratio remains substantially constant for a prescribed number of segments.
In the drawings:
FIG. 1 is a pictorial of a sequence of image frames illustrating an algorithm for indicating the presence of image fading.
FIG. 2 is a block diagram of a video encoder including motion processing and a system in accordance with the principles of the present invention.
FIG. 3 is a flowchart of an algorithm for detecting a fading image.
FIG. 4 illustrates a sequence of line-interlaced frames each containing odd and even image fields which may be processed by the fading detection system discussed in connection with FIG. 1.
FIG. 1 shows a sequence of three adjacent image frames, which in this example contain non-interlaced lines. The following discussion also applies to a sequence of line-interlaced frames, each containing odd and even fields, as will be discussed subsequently. Frame 3 is a current frame which is in the process of being coded, e.g., by the apparatus of FIG. 2 as will be discussed. Frame 2 immediately precedes frame 1 in time, and frame 3 immediately precedes frame 2 in time. Each frame is partitioned into an identical arrangement of 12 horizontal segments designated "A" through "L" in this example. Vertical segments, blocks or other geometric arrangements also may be used. Each frame contains a plurality of image pixels P as known. The number of segments is not critical as long as enough segments, or regions, of an image are provided to reliably indicate the occurrence of image fading in accordance with the requirements of a given system. In this illustration, frame 3 exhibits the most fading (e.g., from a normal image to black), and frame 1 exhibits the least fading in the three frame sequence.
It has been determined that an analysis of three consecutive frames is sufficient to indicate the presence of image fading. Three frame storage devices which may be needed in this regard may be associated with video encoders of the type to be discussed in connection with FIG. 2, or the frame storage devices may be provided independently. Multiple frame storage devices are often found in video encoders for preprocessing operations such as detelecine or prefiltering operations. The following explanation is given with respect to an analysis of corresponding segments "A" in each of the three adjacent frames. Frame difference information obtained from this analysis is used to produce a ratio SA. This value of this ratio, when compared to the similarly computed ratios for all the other segments, indicates whether or not image fading is present. The procedure described below is illustrated by the flowchart of FIG. 3. The FIG. 3 flowchart encompasses the processing of line-interlaced frames as will be discussed.
As a first example, consider a fade-to-black situation as follows
______________________________________Frame No. 1 2 3Fading sequence 1 a a2______________________________________
A fading factor "a" has an arbitrary value but is less than unity since it indicates the reduction in pixel or segment magnitude from one frame to the next as the image fades from normal to black. In practice, the change in luminance level is a good indicator of the amount of change in pixel intensity during fading. This fading sequence (1 . . . a . . . a2) is not predetermined, but is shown as one example of how fading may occur from frame 1 (unity fading factor, no fading), through frame 2 to frame 3 (maximum fading). Thus the amount of fading increases in the progression from frame 1 through frame 2 to (current) frame 3.
Referring to FIG. 1, for corresponding segments A in frames 1 and 2, the difference between the intensity values of similarly located pixels in corresponding segments of adjacent frames is determined. The absolute value of each pixel difference is then obtained. These absolute values are then summed, producing a frame difference value DIFF1-2. A similar procedure is used with respect to frames 2 and 3 to obtain a frame difference value DIFF2-3. Finally, the ratio of frame difference value DIFF2-3 to frame difference value DIFF1-2 is obtained. It can be shown that the frame difference ratio for this example is equal to a value which corresponds to value "a" in the fading sequence illustrated above. This difference ratio for segments "A" among frames 1, 2 and 3 is designated SA. A similar procedure is used with respect to remaining segments B-L to obtain difference ratios SB -SL.
Image fading is indicated if computed frame difference ratio value "a" remains substantially constant among image segments A-L. However, it is not necessary for the ratio value to remain substantially equal for all segments. In accordance with the requirements of a given system, it may be sufficient for a predetermined number of segments, e.g., 80%-90%, or a majority, to exhibit equal ratio values. This number may be a parameter of the algorithm, along with a threshold value for deciding whether or not a computed ratio is equal to value "a." This predetermined number also accounts for the possibility that image motion may be present during the fading sequence.
For example, when there is much motion during a fade and only 40%-50% of the segments have the same ratio value, the fading detection system will declare that no fading has occurred. In this case, even if fading is occurring, it is considered undesirable for the encoder to zero the motion vectors since a large percentage of the blocks being coded may have motion. Zeroed motion vectors will not be used in this case, since the use of non-zero motion vectors will be beneficial. A frame difference ratio value of zero or approximately zero, for example, is an indication that fading has not occurred. More specifically, if the frame difference value DIFF for any single segment (or a predetermined number of segments) is zero or close to zero, the algorithm may declare a condition of no fading. This criterion will prevent a still or nearly still image from being declared a fading image.
A fade from black to a normal scene is the reverse of the fade-to-black process and would be handled similarly. For example, consider a fade from black to normal image situation where the fading factor changes from "a" to "b" over the course of a frame sequence, as follows.
______________________________________Frame No. 1 2 3 4Fading sequence a a2 a2 b a2 b2______________________________________
In this example factor "a" is less than unity, and factor "b" is less than unity but greater than factor "a". As in the other example, this fading sequence is not predetermined, but illustrates another example of how fading may occur from frame 1 (fading begins), through frames 2 and 3 to flame 4 (fading to a normal image completed). Difference values DIFF1-2 and DIFF2-3 are determined for each segment and frame difference ratios SA -SL are computed, as discussed above. It can be shown that the difference ratio for this example is equal to a value "a(1-b)/(1-a)". Image fading is indicated if this value remains substantially constant among all or a predetermined number of image segments A-L, as discussed above. Analogous results will be obtained in the case of a fade from normal to black where the fading factor changes.
To simplify the drawing, FIG. 1 illustrates the fade detection process using three image frames. However, in a line interlaced system, each frame is constituted by an odd numbered image field and an even numbered image field, as known. In such case the previously described difference ratios would be computed separately for the sequence of odd and even fields. For example, the odd field difference ratio SA(odd) would be computed as ##EQU1## where DIFF2-3(odd) is the sum of absolute pixel differences between the odd field of frame 2 and the odd field of frame 3. DIFF1-2(odd) is the sum of the absolute pixel differences between the odd field of frame 1 and the odd field of frame 2. A comparison is made of SA(odd) with SB(odd) . . . up to SL(odd) to determine whether the ratios remain substantially constant for a predetermined number of segments. Similarly, the even field difference ratios are computed, and SA(even) is compared with SB(even) . . . up to SL(even). This separate field comparison enables the system to detect fading in field boundaries.
FIG. 4 shows a sequence of line-interlaced frames 1, 2 and 3 each constituted by an odd field and an even field. In this example the fading factor for the sequence of odd fields changes from a1 to a2 and so on, while the fading factor of the sequence of even fields changes by a different factor, from b1 to b2 and so on. In this type of field fading, the odd and even difference ratios (SA . . . ) are calculated and compared separately.
As a result of the described algorithm, a control signal is developed to indicate whether or not image fading has occurred. This control signal is advantageously used by a video signal encoder to alter signal processing as will be described in connection with the MPEG compatible video (television) signal processing and data compression system shown in FIG. 2.
In the example of FIG. 1, corresponding frame/field segments are subjected to processing in which the sum of the absolute values of pixel differences between co-located pixels in adjacent fields is produced..Other types of processing may be used, depending on the requirements of a given system. For example, for corresponding segments in adjacent fields, the absolute value of the pixels in each segment could be obtained, then summed for each segment. Afterwards the field difference of the sums would be determined. The difference (in contrast to the absolute difference) between corresponding co-located pixels in adjacent similar parity odd or even fields can also be used in the fading detection operation.
In the system of FIG. 2, an input MPEG-compatible digital datastream represents image pixel data of a current image frame to be processed. The input data is stored by an input frame buffer 20. Data read out from buffer 20 is conveyed via a differencing network 22 and a multiplexer 24 to a Discrete Cosine Transform (DCT) unit 26. The operation of units 22 and 24 will be discussed subsequently in connection with motion processing aspects of the system of FIG. 2. The discrete cosine transform performed by unit 26 is a well known technique for effectively and efficiently reducing spatial redundancy of a video signal by transforming an input time domain signal into coefficients representing a discrete frequency spectrum. Each transform coefficient represents a block of 8×8 pixels, for example.
DCT transform coefficients from unit 26 are quantized by unit 28, variable length coded by unit 30, and stored by an output rate buffer 32 before being subjected to forward error correction (FEC) processing by unit 36. The content (fullness) of buffer 32 is monitored by a rate controller 34, which provides an output control signal that is used to adaptively alter quantization parameters (e.g., quantization step size) of quantizer 28. By this mechanism the datastream bit rate is controlled so that the average input rate to buffer 32 is substantially constant. Buffer underflow and overflow are largely prevented, and a substantially constant buffer output bit rate is achieved. The output signal from processor 36 is appropriately processed, e.g., by filtering and modulating, before being conveyed to an output channel.
Blocks 40-54 of FIG. 2 form a known arrangement of a DPCM motion estimation/compensation encoder together with blocks 22-28. A system of this type is described in the Ang et al. article mentioned previously, Video Compression Makes Big Gains. In accordance with the principles of the present invention, an image fading detector 70 is included for generating a Fade Indicator signal. This signal is generated after an analysis of a sequence of adjacent frames as discussed in connection with FIG. 1, and is applied to control inputs of mode control 54 and motion estimator 48. In addition, motion vectors MV developed by motion estimator 48 are applied to a control input of mode processor 54.
The motion processor operates in either an intra-frame coding mode, or in a predictive inter-frame coding mode. In intra-frame coding, the video signal itself is coded. In inter-frame coding, what is coded is a residual image signal which represents the difference between the current frame and the prediction of the current frame. When the prediction is good, inter-frame coding is used and is usually preferred because less frame information needs to be coded. With predictive coding, the value of a pixel is predicted based on its history. The predicted value is subtracted from the current value of the pixel to produce an error, or residual, which is coded and transmitted. A receiver adds the received residual to its own prediction to obtain the correct current pixel value. The discussion which follows assumes predictive inter-frame coding is used.
In the motion processing operation, quantized residual image data from unit 28 in inversely quantized by unit 40 and inversely DCT transformed by unit 42 before being applied to an additive combiner 44. Another input of combiner 44 receives a predicted image which is developed as will be discussed. A reconstructed frame image results from combining the residual image and predicted image inputs to combiner 44. Reconstruction is needed because inter-frame processing uses predictive coding, which requires the encoder to track the behavior of the decoder to prevent the decoder's reconstructed image from diverging from the original input image. The reconstructed image is stored by a frame memory 46. Motion estimator 48 receives a current frame input from the output of input buffer 20, and a reconstructed frame input from frame memory 46. Using a search window, motion estimator 48 compares each (e.g., 8×8) pixel block of the current frame with pixel blocks of the prior reconstructed image from memory 46. For each pixel block a motion vector (MV) is generated to indicate the relative locations of (offset between) the best match between blocks in the current and reconstructed images. The motion vectors are coded by unit 30 and sent to the receiver.
The motion vectors from unit 48 are also provided to motion compensation unit 50, which develops predicted (position adjusted) blocks from memory 46 as a function of respectively associated motion vectors from estimator 48. This operation produces a predicted image comprising (e.g., 8×8) motion compensated blocks from the reconstructed frame. The difference between the predicted image and the current input block, the residual image, is developed by subtractive combiner 22. The residual image is transformed, quantized, coded and conveyed to a receiver/decoder as discussed previously. The decoder performs the inverse of the coder functions. In an inter-frame operating mode, the decoder uses the motion vectors extracted from a variable length decoder to provide the locations of the predicted pixel blocks.
The encoder of FIG. 2 also includes a pair of multiplexers (MUXs) 24 and 52, and an associated inter/intra frame mode processor 54. Each MUX has a pair of inputs for receiving input signals to be switched, labeled "0" and "1." The 0 input of MUX 24 receives the current input signal, and the 0 input of MUX 52 receives a fixed bias level, e.g., a "0" logic level. The "1" inputs of MUXs 24 and 52 respectively receive the residual image signal and the predicted image signal. Mode processor 54 responds to motion vectors (MV) for determining whether inter-frame or intra-frame processing should be used. If the motion compensated prediction is good, mode processor 54 sends a signal to control inputs of MUXs 24 and 52 for enabling inter-frame processing. In this case MUX 52 passes the predicted image signal from unit 50 to adder 44, and MUX 24 is enabled to pass the residual image signal to DCT unit 26 and following units. Conversely, if the prediction is not good, the signal from mode processor 54 enables intra-frame processing. In this case the current frame information itself rather than the residual image signal is passed to DCT unit 24 and subsequent coding units, and the predicted image signal is decoupled from adder 44. The decision as to whether or not a prediction is good is based on a variety of factors which vary in accordance with the requirements of a particular system, such as image quality and the number of bits required.
The signal sent from mode processor 54 to DCT unit 26 instructs unit 26 to either encode the motion predicted error signal (inter-coding), or to encode a macroblock of the original image without motion estimated prediction (intra-coding). Mode processor 54 essentially gives DCT unit 26 a block to process. The inter/intra coding decision is based on efficiency of coding, which may be determined in many ways. Basically, this determination is a function of the number of bits needed to code a macroblock at a given level of quality, and the level of quality at a given number of coding bits. Motion vectors indicate which image region to derive the difference signal from. Intra-coding may be used as a default coding decision if equal coding efficiency is indicated as between intra- and inter-coding. A predetermined threshold may be used in the decision to default to intra-coding.
Mode processor 54 also receives a Fade Indicator control signal from fade detector 70, which operates as explained in connection with FIG. 1. Fade detector 70 may include three frame storage devices for storing the three-frame sequence described previously, as well as digital signal processing circuits for differencing similarly located segment pixels in adjacent frames, a circuit for obtaining the absolute values of the pixel differences, a circuit for summing the absolute values, and a dividing circuit for obtaining the ratio of the summed differences from adjacent frames. A suitable amount of memory may be provided as needed to perform these functions. These elements have not been shown to simplify the drawing.
Briefly, unit 70 provides an indication of image fading by analyzing a sequence of three adjacent image frames to develop a ratio of the sum of absolute pixel differences over the three frame sequence. If a luminance gradient associated with image fading is detected, the Fade Indicator instructs motion estimator 48 to set the output motion vector values to zero (i.e., to coordinates 0,0) for all frames for which fading is detected. Zeroing motion vectors may be accomplished by known techniques, such as by the motion estimator generating variable length codes representative of no spatial displacement. Variations of this result may be used in hierarchical motion estimator systems that set motion vectors at a particular level such as coarse, fine, or half-pixel. The modified output of motion estimator 48 is processed and evaluated, as known, by macroblock decision networks in mode processor 54. The decision circuits preferably should be biased slightly toward an intra-frame coding decision. Although inter-frame coding would use about the same number or slightly fewer coding bits as intra-frame coding, inter-coding would produce more visible coding artifacts.
When the zeroed (0,0) motion vectors are not found to be a good match, mode processor 54 will cause the system to operate in an intra-frame coding mode. However, when zeroed motion vectors are a good match, which is typical for image fading with little or no motion, estimator 48 will produce a good result and unit 54 will instruct the system to perform inter-frame coding.
It has been determined that during image fading, frame-to-frame amplitude (luminance) change is typically more significant than motion related change. Assigning a zero (0,0) value to motion vectors has been found to produce very good results in the presence of fading. Zeroing the motion vectors in the presence of fading also advantageously reduces the number of bits needed for coding, and may be used to alter the coding procedure of a given system, e.g., between inter- and intra-frame coding. Motion vectors of a moving scene are not coded when fading is detected and it is determined that only a small percentage of the image contains motion, e.g., 10% or less, and that frame-to-frame luminance changes do not give rise to motion vectors.
Motion processing systems that are likely to be found in consumer products may not be sophisticated enough to distinguish between true motion, and an image intensity gradient such as a luminance gradient over several frames. These systems may perceive such gradients as "motion" and generate false motion vectors. Coding efficiency will improve when the motion processing system is informed that such a perceived "motion vector" actually has a (0,0) value as discussed above, whereby coding bits are not wasted on false motion vectors.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4779131 *||Jul 25, 1986||Oct 18, 1988||Sony Corporation||Apparatus for detecting television image movement|
|US5162907 *||Sep 18, 1991||Nov 10, 1992||Sony Broadcast & Communications Limited||Motion dependent video signal processing|
|US5181111 *||Jul 24, 1991||Jan 19, 1993||Sony Broadcast & Communications Limited||Video signal processing apparatus and method|
|US5200820 *||Apr 26, 1991||Apr 6, 1993||Bell Communications Research, Inc.||Block-matching motion estimator for video coder|
|US5245436 *||Feb 14, 1992||Sep 14, 1993||Intel Corporation||Method and apparatus for detecting fades in digital video sequences|
|US5363146 *||Feb 17, 1993||Nov 8, 1994||Sony United Kingdom Ltd.||Motion compensated image processing|
|1||*||Video Compression Makes Big Gains, by Ang et al., IEEE Spectrum, Oct. 1991.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5691771 *||Dec 26, 1995||Nov 25, 1997||Sony Corporation||Processing of redundant fields in a moving picture to achieve synchronized system operation|
|US5719642 *||May 7, 1996||Feb 17, 1998||National Science Council Of R.O.C.||Full-search block matching motion estimation processor|
|US5758091 *||Aug 10, 1995||May 26, 1998||Intel Corporation||Method and apparatus for adjusting video data to limit the effects of automatic gain control on motion estimation video coders|
|US5767923 *||Jun 7, 1996||Jun 16, 1998||Electronic Data Systems Corporation||Method and system for detecting cuts in a video signal|
|US5771316 *||Dec 26, 1995||Jun 23, 1998||C-Cube Microsystems||Fade detection|
|US5778108 *||Jun 7, 1996||Jul 7, 1998||Electronic Data Systems Corporation||Method and system for detecting transitional markers such as uniform fields in a video signal|
|US5859668 *||Nov 27, 1996||Jan 12, 1999||Sharp Kabushiki Kaisha||Prediction mode selecting device in moving image coder|
|US5920360 *||Jun 7, 1996||Jul 6, 1999||Electronic Data Systems Corporation||Method and system for detecting fade transitions in a video signal|
|US5959697 *||Jun 7, 1996||Sep 28, 1999||Electronic Data Systems Corporation||Method and system for detecting dissolve transitions in a video signal|
|US6005622 *||Aug 11, 1997||Dec 21, 1999||At&T Corp||Video coder providing implicit or explicit prediction for image coding and intra coding of video|
|US6014183 *||Aug 6, 1997||Jan 11, 2000||Imagine Products, Inc.||Method and apparatus for detecting scene changes in a digital video stream|
|US6037986 *||Jul 16, 1996||Mar 14, 2000||Divicom Inc.||Video preprocessing method and apparatus with selective filtering based on motion detection|
|US6061471 *||Dec 18, 1997||May 9, 2000||Electronic Data Systems Corporation||Method and system for detecting uniform images in video signal|
|US6148035 *||Nov 30, 1999||Nov 14, 2000||Sony Corporation||Processing of redundant fields in a moving picture to achieve synchronized system operation|
|US6317519 *||Feb 9, 1998||Nov 13, 2001||Thomson Licensing S.A.||Process and device for coding by luminance estimation|
|US6341144||Aug 11, 1997||Jan 22, 2002||At&T Corp.||Video coder providing implicit coefficient prediction and scan adaptation for image coding and intra coding of video|
|US6501795 *||Aug 8, 1997||Dec 31, 2002||Canon Kabushiki Kaisha||Adaptable intraframe/interframe motion video encoder|
|US6556625||Aug 31, 2001||Apr 29, 2003||At&T Corp.||Video coder providing implicit coefficient prediction and scan adaptation for image coding and intra coding of video|
|US6560366 *||Dec 12, 1996||May 6, 2003||Paul Gordon Wilkins||Method for analyzing the content of a video signal|
|US6991386 *||Jun 30, 2004||Jan 31, 2006||Fuji Photo Film Co., Ltd.||Image forming apparatus and image forming method|
|US7092445||Dec 31, 2002||Aug 15, 2006||At&T Corp.||Video coder providing implicit coefficient prediction and scan adaptation for image coding and intra coding of video|
|US7281871||Oct 7, 2005||Oct 16, 2007||Fujifilm Corporation||Image forming apparatus and image forming method|
|US7397856 *||Apr 9, 2001||Jul 8, 2008||Siemens Aktiengesellschaft||Method and apparatus for storing and processing picture information of temporally successive pictures|
|US7453938||Mar 30, 2004||Nov 18, 2008||Apple Inc.||Target bitrate estimator, picture activity and buffer management in rate control for video coder|
|US7492820||Mar 30, 2004||Feb 17, 2009||Apple Inc.||Rate control for video coder employing adaptive linear regression bits modeling|
|US7519229||Mar 30, 2004||Apr 14, 2009||Apple, Inc.||Video coding system providing separate coding chains for dynamically selected small-size or full-size playback|
|US7623576 *||Feb 26, 1999||Nov 24, 2009||Stmicroelectronics Asia Pacific Pte. Ltd.||Method and apparatus for interlaced/non-interlaced frame determination, repeat-field identification and scene-change detection|
|US7646809||Apr 18, 2005||Jan 12, 2010||At&T Intellectual Property Ii, Lp.||System and method of providing directional information for direct prediction|
|US7860324||Mar 6, 2009||Dec 28, 2010||Apple Inc.||Video coding system providing separate coding chains for dynamically selected small-size or full-size playback|
|US7869502||Aug 31, 2005||Jan 11, 2011||At&T Intellectual Property Ii, L.P.|
|US7869503||Jan 11, 2011||Apple Inc.||Rate and quality controller for H.264/AVC video coder and scene analyzer therefor|
|US7873225||Jan 18, 2011||Apple Inc.||Video coding system providing separate coding chains for dynamically selected small-size or full-size playback|
|US7889792||Dec 24, 2003||Feb 15, 2011||Apple Inc.||Method and system for video encoding using a variable number of B frames|
|US7974346||Jun 29, 2005||Jul 5, 2011||AT&T Intellectual II, L.P.||System and method for generating video data for implicit coefficient prediction decoding|
|US7986731||Mar 30, 2004||Jul 26, 2011||Apple Inc.||H.264/AVC coder incorporating rate and quality controller|
|US8036267||Dec 23, 2008||Oct 11, 2011||Apple, Inc.||Rate control for video coder employing adaptive linear regression bits modeling|
|US8107539 *||Jul 15, 2003||Jan 31, 2012||Nokia Corporation||Method for error concealment in video sequences|
|US8111929 *||Feb 7, 2012||Interra Systems Inc.||Method and system to detect and quantify blockiness in video files|
|US8311111||Dec 5, 2008||Nov 13, 2012||Google Inc.||System and method for decoding using parallel processing|
|US8325796||Dec 4, 2012||Google Inc.||System and method for video coding using adaptive segmentation|
|US8326075||Dec 4, 2012||Google Inc.||System and method for video encoding using adaptive loop filter|
|US8582656||Apr 13, 2007||Nov 12, 2013||Apple Inc.||Method and system for video encoding and decoding|
|US8593572 *||Oct 8, 2008||Nov 26, 2013||Csr Technology Inc.||Video signal motion detection|
|US8619874||Apr 13, 2007||Dec 31, 2013||Apple Inc.||Method and system for video encoding and decoding|
|US8625665||Nov 30, 2010||Jan 7, 2014||At&T Intellectual Property Ii, L.P.|
|US8780971||Apr 7, 2011||Jul 15, 2014||Google, Inc.||System and method of encoding using selectable loop filters|
|US8780996||Apr 7, 2011||Jul 15, 2014||Google, Inc.||System and method for encoding and decoding video data|
|US8781004||Apr 7, 2011||Jul 15, 2014||Google Inc.||System and method for encoding video using variable loop filter|
|US8885706||Sep 16, 2011||Nov 11, 2014||Google Inc.||Apparatus and methodology for a video codec system with noise reduction capability|
|US8897591||Nov 29, 2012||Nov 25, 2014||Google Inc.||Method and apparatus for video coding using adaptive loop filter|
|US8958481||Nov 8, 2013||Feb 17, 2015||Apple Inc.||Method and system for video encoding and decoding|
|US9131073||Mar 2, 2012||Sep 8, 2015||Google Inc.||Motion estimation aided noise reduction|
|US9154799||Apr 7, 2011||Oct 6, 2015||Google Inc.||Encoding and decoding motion via image segmentation|
|US9262670||Feb 10, 2012||Feb 16, 2016||Google Inc.||Adaptive region of interest|
|US20030059119 *||Apr 9, 2001||Mar 27, 2003||Ralf Buschmann||Method and device for storing and processing image information of temporally successive images|
|US20040139462 *||Jul 15, 2003||Jul 15, 2004||Nokia Corporation||Method for error concealment in video sequences|
|US20050002070 *||Jun 30, 2004||Jan 6, 2005||Fuji Photo Film Co., Ltd.||Image forming apparatus and image forming method|
|US20050147167 *||Dec 24, 2003||Jul 7, 2005||Adriana Dumitras||Method and system for video encoding using a variable number of B frames|
|US20050175092 *||Mar 30, 2004||Aug 11, 2005||Atul Puri||H.264/AVC coder incorporating rate and quality controller|
|US20050175093 *||Mar 30, 2004||Aug 11, 2005||Haskell Barin G.||Target bitrate estimator, picture activity and buffer management in rate control for video coder|
|US20050180502 *||Mar 30, 2004||Aug 18, 2005||Atul Puri||Rate control for video coder employing adaptive linear regression bits modeling|
|US20050226513 *||Mar 30, 2004||Oct 13, 2005||Wallace Gregory K|
|US20050286629 *||Jun 25, 2004||Dec 29, 2005||Adriana Dumitras||Coding of scene cuts in video sequences using non-reference frames|
|US20070248162 *||Aug 31, 2005||Oct 25, 2007||Haskell Barin Geoffry|
|US20080198927 *||Dec 21, 2007||Aug 21, 2008||Tandberg Television Asa||Weighted prediction video encoding|
|US20080253461 *||Apr 13, 2007||Oct 16, 2008||Apple Inc.||Method and system for video encoding and decoding|
|US20080253463 *||Apr 13, 2007||Oct 16, 2008||Apple Inc.||Method and system for video encoding and decoding|
|US20090103610 *||Dec 23, 2008||Apr 23, 2009||Apple Inc.||Rate control for video coder employing adaptive linear regression bits modeling|
|US20090168898 *||Mar 6, 2009||Jul 2, 2009||Apple Computer, Inc.|
|US20090169124 *||Mar 6, 2009||Jul 2, 2009||Apple Computer, Inc.|
|US20090190030 *||Oct 8, 2008||Jul 30, 2009||Zoran Corporation||Video signal motion detection|
|US20090220152 *||Feb 29, 2008||Sep 3, 2009||Praveen Kumar Tiwari||Method and system to detect and quantify blockiness in video files|
|US20100061444 *||Dec 5, 2008||Mar 11, 2010||On2 Technologies Inc.||System and method for video encoding using adaptive segmentation|
|US20100061455 *||Mar 11, 2010||On2 Technologies Inc.||System and method for decoding using parallel processing|
|US20100061645 *||Mar 11, 2010||On2 Technologies Inc.||System and method for video encoding using adaptive loop filter|
|US20110075739 *||Mar 31, 2011||At&T Intellectual Property Ii, L.P.||Video Coder Providing Implicit Coefficient Prediction and Scan Adaptation for Image Coding and Intra Coding of Video|
|CN1127266C *||Feb 10, 1998||Nov 5, 2003||汤姆森多媒体公司||Process and device for coding by luminance estimation|
|U.S. Classification||348/700, 375/E07.256, 375/E07.263, 375/E07.224, 375/E07.148, 348/E05.067, 375/E07.211, 348/594, 348/E05.065|
|International Classification||G06T9/00, H04N11/04, H04N5/14, H04N7/32, H04N7/26, H04N7/36, H04N7/50, H03M7/36|
|Cooperative Classification||H04N19/503, H04N5/144, H04N19/61, H04N19/107, H04N19/51, H04N5/147|
|European Classification||H04N7/26A4C2, H04N5/14M, H04N7/36C, H04N7/50R, H04N7/50, H04N7/36D|
|Oct 26, 1994||AS||Assignment|
Owner name: THOMSON CONSUMER ELECTRONICS, INC., INDIANA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:POLIT, PETER PAUL;REEL/FRAME:007208/0476
Effective date: 19941020
|Dec 15, 1999||FPAY||Fee payment|
Year of fee payment: 4
|Feb 27, 2004||FPAY||Fee payment|
Year of fee payment: 8
|Mar 7, 2008||FPAY||Fee payment|
Year of fee payment: 12